Residential mobility in pregnancy and potential exposure misclassification of air pollution, temperature, and greenness

Introduction: Epidemiological studies commonly use residential addresses at birth to estimate exposures throughout pregnancy, ignoring residential mobility. Lack of consideration for residential mobility during pregnancy might lead to exposure misclassification that should be addressed in environmental epidemiology. Methods: We investigated potential exposure misclassification from estimating exposure during pregnancy by residence at delivery utilizing a prospective cohort of pregnant women in New York, United States (n = 1899; 2016-2019). We calculated exposure during pregnancy corresponding to each address for fine particles (PM2.5), temperature, and greenness (Enhanced Vegetation Index [EVI]). Results: Twenty-two percent of participants moved at least once during pregnancy; 82.3% of movers changed residences during the second or third trimesters. Participants with better health, lower parity, and higher socioeconomic status were more likely to move. Exposures based on address at delivery rather than residential history overestimated exposure for PM2.5 (exposure error: range -5.7 to 4.6 mu g/m(3), average -0.6 mu g/m(3)) and EVI (range -0.305 to 0.307, average -0.013), but not temperature. Overestimations were significantly larger for mothers with higher socioeconomic status. Our findings indicate that the error for prenatal exposure can occur when residential mobility is not considered and is disproportional by maternal characteristics. Conclusions: Epidemiological studies should consider residential mobility in exposure assessments based on geolocation when possible, and results based on mother's residence at birth should be interpreted with understanding of potential differential exposure misclassification.